Users losing data when migrating from Internal directory to AD

[Kamil Beer]

Hey guys, I was hoping you might have an idea on what I did wrong in this Internal directory -> Crowd migration...

We have 500 internal users and my goal is to start using AD for user access. So, I connected my AD to Crowd. Then, I connected Crowd to Jira. The AD users represented the same people as the accounts in the internal directory.

I gave the internal users the same usernames they had in AD, so new accounts weren't created in the internal directory (duplicated) - and put Crowd as first in the directory order so they can log through AD.

The only accounts that appeared from that point were the ones from Crowd (which is correct), but the users lost all their data they had when they were logging from the internal directory. When they logged in through Crowd (AD), they had no assigned issues, dashboards and so on.

My goal is to start using AD for user access, but without the users losing the data they had in the Internal directory (assignees and so on).

What did I do wrong?

[Thomas Deiler]

Dear @Kamil Beer ,

this sound like the mapping got lost. Probably too late for you, but for other readers:

Never do such critical operations on production environments without verifying first on a staging environment.

So  how this with the user works. Following case:

• Jira is attached to AD. (LDAP)
• users are assigned to project roles
• users are deleted from AD
• the projects roles are not removed, because deleting an AD user is not causing an action in jira
• to revoke the lost user relation create a new "synthetic" user in Jira Internal directory
• important: just the login has to be identically
• The relation gets visible again an can be deleted

Also good to know: if you have identical logins of different user directories the first is taken which will shadow the logins afterwards.

Now coming to your problem: Hopefully you didn't delete the internal users. Push this internal user directory back on top and check if the users have kept their settings.

First, connect AD directly to Jira without Crowd and check if this works. Be sure that the AD login is mapped correctly.

Use the same LDAP settings when you connect AD to Crowd.

If all is fine add Crowd to Jira leaving AD where it is, just the priority of the user directories will change. (for fallback)

If you are sure disable AD (not delete).

Let me know if you need more support.

So long

Thomas